How progressives should respond to another polling failure

How progressives should respond to another polling failure

By Ryan Cooper.

November 5, 2020.

The presidential election has not yet been called, but Joe Biden seems to have an insurmountable lead in enough states to secure victory in the Electoral College. Nevertheless, it's clear that the polls were seriously off in many states. Biden was predicted to win easily, with comfortable margins in some states that ended up nearly even and the possibility of picking up several more that he ended up losing decisively. Similarly, several Democratic Senate candidates performed well below their expected numbers.

As a result, many progressives are understandably frustrated about the second consecutive polling whiff in a presidential election year. Staggering amounts of money were wasted on campaigns that were not close at all. About $90 million went into Amy McGrath's doomed vanity campaign for Senate in Kentucky, where she lost by over 20 points — but another $108.9 million went into Jaime Harrison's campaign against Lindsey Graham in South Carolina, which polls showed to be close but ended up being a 10-point loss. Something is very wrong with pollsters' methods.

What is to be done? I suggest that progressives and Democrats should sink some of their money into local research — organizations that will conduct journalism, interviews, focus groups, listening sessions, and so forth to supplement surveys and polls. That could both provide some more qualitative data to balance out untrustworthy polls, and also build up an institutional presence on the ground that will be valuable in future political fights.

Some might say that Democrats should just run on quality policy, and completely ignore the polls. And it's true, as I have argued many times previously, that Democrats tend to overrate opinion polls in making their decisions (or hide behind them when they don't want to do something that would alienate big donors). Many people do not have well-formed opinions about the issues, and can be convinced one way or another with careful messaging and strong leadership. Polls should never be dispositive, only informative.

But knowledge of public opinion must play a vital role in any kind of modern political campaign. Even in these big spending times, resources are not unlimited. A decent idea of how close races are allows organizations to prioritize their spending, while granular data about locations and demographics allows campaigns to shape their messaging and their turnout efforts. If a candidate doesn't know where he or she is stronger or weaker, or how people are reacting to arguments, it's impossible to fight tactically. Even back in the pre-polling days, political campaigns relied on union leaders, local party bosses, ward heelers, and so on to provide some crude intelligence.

Now, local research can be even more misleading than bad polls. This kind of anecdotal journalism, for instance, is often a complete joke. One must avoid the Salena Zito problem, where a journalist supposedly doing shoe leather reporting miraculously finds that everyone (that is, a bunch of Republican partisans that Zito claims are Democrats) agrees with her pro-Trump politics. More generally, one must avoid the temptation to talk one's book — shading the research to advance one's personal politics, which can happen even unconsciously.

That said, local research can capture political intelligence that even accurate survey data can miss. Structured conversations, knowledge of local relationships, or even just a feel for the local mood can provide valuable information even if it can't be programmed into a spreadsheet. For every Zito article or "Trump supporters who still support Trump" piece in The New York Times, there is work like Michael Lewis on the 2008 financial crisis in Iceland, or Sharon Lerner on "Cancer Alley" in Louisiana. Academic anthropology has developed elaborate tools for conducting ethnographic studies in an informative, fair way. (Incidentally, there are a large number of unemployed or struggling academics who would probably leap at the chance to do quality field work for a steady paycheck.)

This kind of thing is already happening in other contexts. The Federal Reserve, for instance, has access to all the best economic data available — far, far better than any poll. Yet before the pandemic struck, it organized a series of "Fed Listens" tours across the country, where Fed officials had long conversations with local business and community leaders. Ordinarily one would think this would be a silly exercise in public relations, but as Brendan Greeley reports for the Financial Times, it actually had a marked effect on Fed leaders' thinking and policy. They discovered in a visceral, intuitive way that basically no Americans are concerned about inflation, but were very concerned with jobs and incomes. It's partly why the Fed correctly moved to stimulate the economy even before the pandemic struck, and some officials are suggesting the tours should be institutionalized as part of Fed procedure. That type of knowledge and experience can make a good complement to even the best data in the world. "I've been pleasantly surprised at how meaningful they were," said Neel Kashkari, head of the Minneapolis Fed branch.

Progressives donated an absolute avalanche of money in the final stages of this election. Just in the third quarter of 2020, small donors funneled $1.5 billion through the donation portal ActBlue — nearly matching the entire total for 2018. It seems that donations are what millions of progressives are doing to relieve their intense desire to do something to fight the Republicans. And I understand why! But building some local institutions, inside or outside the Democratic Party, would give people something to do that would be more concretely useful than the marginal billionth dollar, and could easily be spun up into actual campaign work. (It seems Biden's abandonment of door knocking was likely a significant factor in his underperformance.)

There is no replacement for accurate polls. But people are going to be suspicious of pollsters for a long time, at least until they stop faceplanting for several elections in a row. In the meantime, Democrats and activist groups alike can divert just a bit of that money firehose and start opening some permanent field operations.